MBMUH: XXIX.] ( 11KMK Al. JOB I IN llll. BPK 1 Kl'M. 



dec-line, tlie shorter waves of these are the first to b< 

 Anguished, the longer ones the last. This, therefore, i- 

 in accordance with what I found to be the case in the 

 -i ;ulual warming of a solid body, in which the long waves 

 pertain to a low temperature, the short ones arising as 

 the temperature ascends (Memoir I.). 



In some cases, however, instead of the disturbing un- 

 dulation giving rise to longer waves, it produces shorter 

 ones, as is shown when a platinum wire is put into a 

 hydrogen flame, or by Tyndall's experiments, in which 

 invisible undulations below the red give rise to the igni- 

 tion of platinum. 



(2d.) If the substance be colored and undecomposable, 

 it will extinguish rays complementary to its own tint. 

 The temperature will rise correspondingly. 



(3d.) If the substance be decomposable, those portions 

 of the radiation presented to it which are of a comple- 

 mentary tint will be extinguished. The force thus dis- 

 appearing will not be expended in establishing vibrations 

 in the arresting particles, but in breaking down the union 

 of those which have arrested them from associated par- 

 ticles. No vibrations, therefore, are originated ; no heat 

 is produced ; there is no lateral conduction. 



In actinic decompositions the effects may be conven- 

 iently divided into two phases: 1st, physical ; 2d, chemi- 

 cal. The physical phase precedes the chemical. It con- 

 sists in a preliminary disturbance of the group of mol- 

 ecules about to be decomposed. Up to a certain point 

 the dislocation taking place may be retraced or reduced, 

 and things brought back to their original condition. 

 But that point once gained, decomposition ensues, and 

 the result is permanent. 



I may perhaps illustrate this by a familiar example. 

 If a sheet of paper be held before a fire, its surface will 



EE 



